Research On Rapid Alignment Detection Of Micro Assembly System

With the development of micro manufacturing technology,MEMS has been widely used in aerospace,national defense industry,modern medicine and biological engineering because of its small size,low energy consumption,stability and reliability.Due to the assembly process is complicated and requires high assembly accuracy,the traditional manual assembly has the problems of low efficiency and low rate of finished products.Machine vision technology has advantages in non-contact,high precision and fast speed.So,it can effectively compensate for the shortcomings of manual assembly and improve production efficiency in the field of micro assembly.In this paper,micro plate part is used as the object,research on the alignment detection technology,including part image processing,assembly feature recognition and imaging system calibration,is carried out based on the second generation micro flexible automated assembly system developed by research group.Also,the template matching method based on contour feature is applied to the recognition of part contraposition feature,which effectively improves the detection efficiency.The main research work is as follows:1.The construction of alignment detection system for automatic assembly of micro flat panel parts is completed based on the basic principle of coaxial alignment,including the hardware selection of micro imaging system,the design of prism position adjustment mechanism,also the selection,installation and debugging of precision displacement table in the assembly execution system.Based on VS2010 platform,programming of image acquisition,processing,recognition and position deviation calculation is completed by using MIL and Opencv computer vision library.2.The contrast enhancement,filtering and preprocessing method of dryness and thresholding image are systematically introduced.According to the imaging characteristics of micro flat panel parts,a specific algorithm combination is selected to process the part images.The template matching method based on contour feature is applied to feature recognition of parts,so the assembly parts can be quickly and accurately located.3.The model of coaxial contraposition detection system imaging is analyzed,the camera parameters and micro rotary center is calibrated.The influence of prism attitude error on detection accuracy is studied,and the attitude of prism is corrected.The part relocate control information conversion model is studied,and the compensation of camera mounting error and cross platform quadrature error is added to ensure the accuracy of position detection.4.The effectiveness of the image processing algorithm and contour feature matching method based on template in the application of parts counterpoint in feature recognition is verified through the typical micro structure assembly experiments.The standard axle bore was assembled to verify the comprehensive accuracy of contraposition assembly system.